AI Won't Replace Your Kid. The Kid Next to Them Might.

Every week someone asks me some version of the same question. A friend during game night. An acquaintance during a dinner. My stepmom during lunch.

Not product managers. Not engineers. Parents. Grandparents.

“Should I be worried about my kid?”

Their son just declared Computer Science as their major, and they are not sure that means what it used to mean. Their grandson is a Junior worried about their major choice. Their daughter is about to graduate into a job market that shifted underneath her while she was still in class. A friend’s kid can’t find an internship and nobody can explain why.

They’ve read the headlines. They’ve seen the layoffs. They know enough to be worried but not enough to know what to worry about. They want someone who works in this space to tell them it’s going to be okay.

I can’t tell them that. Not the way they want to hear it.

But I can tell them something more useful. And more honest.

The honest answer

Here’s what I’d actually tell them.

The worry is real. According to the NY Fed, recent computer science graduates face roughly 7% unemployment, well above the 4.3% national rate. SignalFire’s State of Tech Talent report found that Big Tech cut new grad hiring 25% since 2023. New grads now account for just 7% of hires.

The job market your kid is walking into is not the one you walked into. That part is true.

And it is not just entry-level. Companies are making a bet. That same SignalFire report found mid-level tech hiring is up 27%. The market is voting for experience over speed. That tension is real, and it matters for what comes next.

The edge nobody’s talking about

But let’s look at this more positively. Here’s what most of the headlines seem to miss.

This generation has an edge that no one older than them will ever have again. They are “AI Natives.”

They are learning how to think and how to work with AI — at the same time. Not retrofitting. Not unlearning 20 years of habits to make room for a new tool. Building from the ground up with it already in their hands.

That matters more than people realize.

Their computer science foundations are fresh. Data structures, systems architecture, how things connect and talk to each other. The stuff that gets rusty after a decade in a senior role. They understand what’s underneath the tools in a way that many experienced operators, and even some product managers, honestly don’t. And they have AI layered on top of that foundation from day one.

A senior operator like me spent years developing judgment through repetition. Thousands of decisions, thousands of mistakes, pattern recognition built the slow way. That experience is genuinely valuable. I’ll make that case in an upcoming article.

But the new grad? They are building that same muscle at a speed that wasn’t possible two years ago.

They can prototype in an afternoon what used to take a team a quarter. They can test five hypotheses before lunch. They move faster than anyone else in this space right now, and it’s not close — by a long shot.

The CEO of Google DeepMind said it recently: the opportunity is enormous for people who immerse themselves in these tools and apply them to domains nobody has tried yet. He’s right. But there’s a condition he didn’t name.

The fork

In my last article I said: “don’t defer your intelligence to the tool.” That was aimed at experienced PMs. But it applies even more to the generation that never worked without one.

And that’s where this gets complicated.

Two students. Same classroom. Same tools. Same access to the same AI. One is building things, breaking them, defining what “good” looks like before the tool runs. The other is pasting prompts, accepting outputs, skipping the thinking.

From the outside they look identical. Same degree. Same resume line. Same graduation date.

They are not the same.

Some Google teams recently changed how they interview product managers. 45 minutes. Build a working prototype in Cursor (a coding tool). Live. No slides. No hypothetical case studies. No “tell me about a time when.” Just: show me.

Either they fail for 45 minutes, or they build in real time while the interviewer watches them succeed. There is no middle ground.

Across frontier AI companies, the interview question has converged to the same three words: show me your setup. The bar moved from talking about building to actually building, in the room, under pressure.

The student who’s been building like that every day compounds into that moment. Every prototype, every failed experiment, every time they defined the outcome before they ran the tool. It all shows up in 45 minutes. After six months of that rhythm, the gap between them and the student who skipped the thinking isn’t a gap anymore. It’s a career.

The student who’s been deferring hits a wall with people watching.

That’s the fork. Same tools. Same degree. Completely different trajectory.

Now think about which one your student is.

Critical thinking was always important. It was just never this visible. AI didn’t change what matters. It made what matters impossible to fake.

What the right side looks like

So what does the right side of that fork actually look like?

It’s not about which AI tool you use. Nobody who’s hiring cares whether you learned on ChatGPT or Claude or Copilot.

They care what you built with it. Which problem you picked.

Some companies are already creating new roles around this. Not “AI Engineer.” Not “Prompt Specialist.” Product Builders. People who can take a problem and ship something real with these tools. The title is the tell.

Build things. Not assignments. Things. Projects you chose, problems that are hard, prototypes you made because you wanted to see if they’d work. I met a UC Santa Cruz student last month who was building a voice agent you can call to help illiterate inmates understand their legal challenges. Nobody assigned that. He picked a problem that mattered to him and used AI to go after it. A portfolio of work like that says more than any transcript.

Here’s the part most people skip.

Before you let the tool run, know what you’re looking for. Define what “good” looks like. Learn to recognize it. Form your hypothesis. Have a position. Then use AI to pressure-test it, build it, break it. The student who does this is building judgment — quickly. The one who opens a prompt with “make me a…” and accepts whatever comes back is not.

That’s the difference between using AI and deferring to it.

And then there’s the part nobody wants to hear.

The soft skills matter more now, not less. When AI can tell you what the data says, the differentiator is the person who can walk into a room, articulate what they think and why, and get six people aligned on a direction. Reading the room. Rallying people. Making the case for the thing you believe in and against the thing you don’t.

That’s critical thinking. Clearly and concisely articulating what you believe and why. Not the loudest voice in the room. The clearest one. It was always the job underneath the job. AI just made it the only part that’s left.

I know what some of you are thinking. My kid is an introvert. My kid isn’t comfortable in social settings. “Develop charisma” isn’t guidance. It’s platitude.

Fair. But here’s the reframe. That 45-minute prototype interview? It doesn’t care how you present. It cares what you built. And critical thinking in an AI world doesn’t mean be the loudest person in the room. It means be clear about what you think and why. That’s a muscle anyone can build.

What to actually ask your kid

So when someone asks me “should I be worried about my kid?” here’s what I actually say.

Ask them WHAT they are building. Not what classes they are taking. What they are building.

Ask them which tools they are using. Are they exploring what their classmates are using? Are they directing AI agents or just accepting what comes back? Are they constantly iterating, flipping back and forth between their thinking and the tool’s output, or are they submitting a prompt and calling it done?

Better yet, give them an idea and say: show me how you’d build it in 30 minutes. You’ll know in the first five minutes which side of the fork they are on.

Then ask the harder questions. What problem are you solving? Why does it matter? Who is it helping? Can you explain why what you built is good — not just that it works, but why it’s the right thing? Why you made the choices you made?

That’s judgment. That’s the muscle. And that’s the thing no AI can build for them.

The question that matters: are you using AI to build your thinking, or to skip it?

The answer you get tells you which side of the fork they are on. And unlike the job market, that’s something they can actually control.

One more thing

Everything I just said about the next generation applies in reverse to the generation above them. Same fork. Different direction. People like me. But that’s the next article.